Campus Event Calendar
Event Entry
What and Who
"Split Snapshots: A New Approach to Time Travel in Storage"
Liuba Shrira
Brandeis University
SWS Colloquium
Liuba Shrira is an Associate Professor in the Computer Science Department at Brandeis University,
and is affiliated with the Computer Science and Artificial Intelligence Laboratory at MIT.
From 1986 to 1997 she was a researcher in the MIT/LCS Programming Methodology Group
joining Brandeis in 1997.
In 2004-2005 she was a visiting researcher at Microsoft Research, Cambridge, UK.
Her research interests span aspects of design and implementation of distributed systems and especially storage systems.
This includes fault-tolerance, availability and performance issues. Her recent focus is on long-lived transactional
storage, time travel (in storage), software upgrades, byzantine faults, and support for collaborative access to
long-lived objects.
AG 1, AG 2, AG 3, AG 4, AG 5, SWS, RG1, RG2
Expert Audience
English
Note: We use this to send email in the morning.
Date, Time and Location
Wednesday, 18 June 2008
10:30
60 Minutes
E1 5
rotunda 6th floor
Saarbrücken
Abstract
Kurzweil says, computers will enable people to live forever and doctors will be doing backup
of your memories by late 2030. This talk is not about that, yet. Instead, the remarkable drop in
disk costs makes it possible and attractive to retain past application states and store them
for a long time for "time-travel". A still open question is how to best organize long-lived past
state storage?
Split snapshots are a recent approach to virtualized past states that is attractive for several reasons.
Split snapshots are persistent, can be taken with high-frequency, and theyare transactionally
consistent. Unmodified database application code can run against them. Like no other approach,
they provide low-cost discriminated garbage collection of unneeded snapshots, a useful feature in
long-lived systems. Lastly, compared to a temporal database, split snapshots are significantly
simpler and more general since they virtualize disk blocks rather than logical records.
Several novel techniques underly split snapshots. An extended recovery invariant allows to
create consistent copy-on-write snapshots without blocking, a new kind of persistent index
provides fast snapshot access, and a new snapshot storage organization incrementally garbage
collects selected copy-on-write snapshots without copying and without creating disk-fragmentation.
Measurements of a prototype system indicate that the new techniques are efficient and scalable,
imposing minimal ($4\%$) performance penalty on a storage system, on expected common workloads.
(Joint work with Ross Shaull and Hao Xu)
Contact
Brigitta Hansen
0681 - 9325200
--email hidden
passcode not visible
logged in users only
Brigitta Hansen, 06/13/2008 16:41 -- Created document.